Indicator Fact Sheet (WHS6+WHS7) Hazardous substances in marine organisms and loads to coastal waters: DDT

Transcription

1 Indicator Fact Sheet (WHS6+WHS7) Hazardous substances in marine organisms and loads to coastal waters: DDT Author: Norman Green, NIVA EEA project manager: Anita Künitzer version Key message All of the significant trends found for Baltic herring were decreasing Estimations indicate an inconsistent but decreasing trend for DDT in mussels from the north-east Atlantic. Estimations indicate decreasing trend was evident for concentrations in herring from the Baltic. Estimations for inputs and concentrations in mussels and fish have been derived from data that are unbalanced, incomplete or both. Estimations indicated considerable variation in concentrations in cod liver. (A) DDT concentrations in blue mussel (Mytilus edulis) from the north-east Atlantic. (B) DDT concentrations in liver of cod (Gadus morhua) in the north-east Atlantic and muscle of herring (Clupea harengus) in the Baltic. % relative to year 1990 level Blue mussel, N.E. Atlantic % relative to year 1990 level Herring muscle - Baltic Cod liver - N.E. Atlantic Figure 1: Change (%) in DDT in (A) concentrations in blue mussels (M. edulis) in the north-east Atlantic (B) concentrations in Atlantic cod (Gadus morhua) in the north-east Atlantic and herring (Clupea harengus) in the Baltic Data source: OSPAR and Helcom. Results and assessment Due to human activities inputs of DDT into European coastal waters have increased especially in the 1970s and early 1980s. The presence of hazardous substances, increased inputs and concentrations of hazardous substances in coastal waters, estuaries, fjords and lagoons may negatively affect the ecosystem quality. European policies aim at reducing the inputs and improving the state of the marine and coastal environment. Inputs and atmospheric deposition of DDT constitute a pressure indicator for marine and coastal water quality. Emission sources and emission patterns differ between hazardous substances. Thus, in determining the results of pollution abatement policies, each substance should also be considered separately. DDT is not on the EU's list of priority substances (2455/2001/EC (EU, 2001)), but is on the OSPAR list 1

2 of chemicals for priority action (OSPAR, 1998). Policy relevance: target or objective for the indicator The objective of indicators is to convey the levels and trends of hazardous substances of inputs and concentrations in European seas. The effect of hazardous substance load, especially years ago, may be detrimental to marine ecosystems. Concentrations of hazardous substances in blue mussels and fish constitute time integrating state indicators for coastal water quality. An advantage to using biota concentrations as indicators, as opposed to using water or sediment, are that they are of direct ecological importance as well as relevant to human health and economical factors due to consumption. Mussels are attached to shallow-water surfaces, thus reflecting exposure at fixed point. The disadvantage of this aspect is that they are restricted to the coastal zone. Fish are exposed to pollution over wider areas and can in some cases reflect offshore conditions. Policy context (relevance of the indicator with reference to specific policy processes) Measures to reduce riverine inputs, direct discharges and atmospheric deposition hazardous substances and to protect the marine environment are being taken as a result of various initiatives on different levels: UN Global Programme of Action for the Protection of the Marine Environment against Land-Based Activities; Dangerous Substances Directive (76/464/EEC); Water Quality Framework Directive (2000/60/EU); OSPAR Convention 1998 for the Protection of the Marine Environment of the north-east Atlantic, Mediterranean Action Plan (MAP); Helsinki Convention 1992 on the Protection of the Marine Environment of the Baltic Sea Area (Helcom); and AMAP, the Arctic Monitoring and Assessment Programme. The target of these initiatives is a substantial reduction of the input of hazardous substances to coastal waters, thereby improving the biological state. However, for the time being there is no specific management target for the indicators. In the OSPAR work continues towards the reduction of discharges, emissions and losses of hazardous substances which could reach the marine environment, to levels that are not harmful to man or nature with the aim of their elimination. The Commission will implement this strategy progressively by making every endeavour to move towards the target of the cessation of discharges, emissions and losses of hazardous substances by the year 2020 (OSPAR 1998). As targets have been formulated for emissions and the current indicator presents input loads into the coastal waters, only an approximate comparison is possible between the measured input and the reduction target aimed at emissions. The ministers at The Fifth International Conference on the Protection of the North Sea in March 2002 stressed that increased efforts are necessary in order to meet the OSPAR target. Helcom has adopted the Recommendation 19/5 in May 2001 for cessation of hazardous substance discharge/emissions by 2020, with the ultimate aim to of achieving concentration in the environment near background values for naturally occurring substances and close to zero for manmade synthetic substances. No target could be found for Medpol. Water quality limits and quality objectives for discharges of DDT are primarily regulated by the Dangerous Substances Directive (76/464/EEC) on general restrictions on emissions, Council Directive 86/280/EEC. DDT (and its metabolites) is included on List I of Council Directive 76/464/EEC (Water pollution by discharges of certain dangerous substances). Environmental context: (scientific soundness and choice and definition of the indicator) 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane (DDT) is a persistent synthetic pesticide. It is degraded primarily to DDE (1,1-dichloro-2,2-bis(chlorophenyl) ethylene) and DDD (1,1,- dichloro-2,2,-bis(p-chlorophenyl) ethane). The use of DDT is banned in most European countries, but is still in use elsewhere. The original pesticide DDT (technical grade) contained 2

3 up to 14 different chemicals, including DDD and DDE. There are both p,p and o,p isomers of all three substances, varying in their biological activity. In most current contexts (and in this Fact Sheet), DDT will denote the sum of DDE and DDD because DDT itself is not frequently measured. DDT is not found naturally, but has over the past years been spread over the entire globe and is now found in all natural waters and all organisms. DDT was originally used as an insecticide, but also affects other invertebrates and vertebrates. Due to its persistence, DDT will bioaccumulate and -magnify in food webs. The perhaps most well-known effect in the marine environment is associated with the interaction of DDT (actually primarily its metabolites) with eggshell-production in birds. Due to their position high in food-chains, piscivorous seabirds and terrestrial predatory birds were most strongly affected. Although other contaminants were probably also involved, DDT contributed to a decline in many bird populations in industrialised areas of the northern hemisphere in the 1960s and 1970s. Along with other organic contaminants, DDTs are suspected of affecting reproduction in populations of seals in the Baltic and polar bear in the European Arctic. All use of DDT was discontinued in west European countries around 1990, although heavy use was banned two decades earlier. DDT still appear to be used in some countries, especially in the third world, due to its high efficiency in combating mosquitoes (and hence malaria). In some areas, e.g. in fruit-growing areas in western Norway, there are recent inputs of DDT, probably due to leaching from buried waste or unused canisters. In addition, there is a continued leaching from soil and river sediments in some areas. Mussels are attached and therefore concentrations of harmful substances reflect the local pollution. Also mussels are abundant, robust and widely monitored for hazardous substances concentrations in soft tissue in a comparable way. Atlantic cod is widely distributed and a commercially important fish species in the north Atlantic. It is a predator and as such it will also to some degree reflect contamination levels in its prey. Herring is commercially and ecologically important species both in the Baltic and the north-east Atlantic. The developed indicator is expressed as the percentage of change of concentrations of hazardous substances in a year compared to 1990, showing directly the success of the reduction measures taken. A more detailed description of the development of the indicators on hazardous substances is under publication (Baan & Groeneveld, in press). Assessment There is no available data for direct and riverine or atmospheric inputs of DDT. Emission reduction measures in industry and political changes in Europe have presumably contributed to the reduction. EU legislation have not set foodstuff limits for DDT in mussels and fish. An index (i.e. the ratio of annual concentration to average concentration) for DDT (measured as sum of DDE and DDD) show decreasing trends at many localities. Of the 78 mussel stations in European coastal waters (only from the north-east Atlantic and Baltic), the median values for only 3 stations in the period exceeded the OSPAR suggested low concentration as ecotoxicological assessment criteria (Table 1, Map 1). Of the 40 fish stations, only 4 exceeded suggested low concentrations (Map 2). Figures 1A and 1B show time trends of concentrations in mussels and fish for each region as averages relative variation in medians in time over locations with sufficient data for the period (see Table 2 and Metadata for more details). The figures give visual indications of the general development; without any formal assessment of statistical significance. The trends are averages over monitoring stations with very different temporal coverage, and also with very different time change patterns. No attempt to weight the stations according to 3

4 representativeness has been made, due to lack of information. However, a visual comparison with Maps 3 and 4 of time development for each time series indicate that Figures 1A and B are reasonable summaries of the development over a majority of the locations included. They indicate an inconsistent but decreasing trend for DDT in mussels from the North-East Atlantic. The level in herring muscle in the Baltic is decreasing. The levels in Atlantic cod have varied considerably. Of the 178 temporal trends (99 for mussels, 79 for fish) analysed only 15 were significant, and 13 were down (Table 2, Maps 3 and 4). Substantial long-term trends might be undetected due to insufficient data. Statistically significant trends were found mainly at locations in estuaries and fjords, which are closer to the sources of land-based diffuse pollution. Concentrations of DDT in blue mussels is low all over Europe. Levels above background has recently only been observed in mussels collected from a site in the Sørfjord, western Norway. The lack of change in DDT concentration in mussels at most sites presumably indicates that the current background level is maintained through ecosystem recycling or inputs through long-range transport. There is no known reason for the increase found at the one site in eastern Iceland. Similarly to blue mussels, the concentrations of DDT in Atlantic cod liver is at background levels in most areas, even in the Baltic. Somewhat elevated concentrations, but below 10 times the background concentration, were found at three locations in Norway. The sites reflected both fresh discharges of DDT (by leaching from containers or waste) and leaching from old deposits (identified by the ratio of DDT to its metabolites). As expected, DDT concentrations in fish has decreased in most areas, including the Baltic and the Scheldt (Rhine estuary). An exception is cod from the Oslofjord and the southeast coast of Iceland. More detailed study of the background data indicate that one out of three dataseries from the same area shows an increase whereas the other two show no change. The increase may simply be a statistical artifact with such a large number of tests. For the Black Sea, there is too little data available to make assessments. The concentration of DDTs in mussels for the single location with data on the Romanian Coast are all in the intermediate range (Map 1), but this may not be representative for the region. 4

5 Sub-indicators: overview of DDT levels and their temporal trends Table 1. Limit values (low/high concentration) for DDT in marine organisms for spatial assessment Name and tissue Latin name Low/ High mg/kg wet Ref. Comment weight Mussels Mytilus sp. 1 Low 0.01 OSPAR, 1999 EAC 2 limit DDE. A BRC 3 of ppm wet weight has been suggested (Green & Knutzen, in press) Mussels Mytilus sp. High 0.10 Taken as 10 time Low Atlantic cod, liver Gadus morhua Low 0.5 OSPAR, 1999 EAC 2 limit DDE. A BRC of 0.2 ppm wet weight has been suggested (Green & Knutzen 1987) Atlantic cod, Gadus morhua High 5.0 Taken as 10 time Low liver Herring, muscle Clupea harengus Low 0.05 Knutzen, 1987 Upper limit to proposed range in background level Herring, muscle Clupea harengus High 0.50 Taken as 10 time Low ( 1 ) Blue mussel (Mytilus edulis) for the north-east Atlantic, Mediterranean mussel (M. galloprovincialis) for the Mediterranean. ( 2 ) EAC = ecotoxicological assessment criteria. ( 3 ) BRC = background reference concentration. ( 4 ) Only along the Romanian coast. Table 2. Spatial variation and temporal trends of DDT concentrations Spatial variation (Map 1 and 2) Time trends (Map 3 and 4) Sea Subindicator Total no. of stations Number over Low Number over High Total no. of stations Number Down Number Up NE Atlantic Mussels Fish Baltic Mussels Fish Mediterran. Mussels 0 0 Fish 0 0 Black Sea 4 Mussels

6 Map 1. DDT (expressed as DDE+DDD) in mussels (Mytilus edulis north-east Atlantic ; M. galloprovincialis and Black Sea, median mg/kg wet weight for (2001 for the Black Sea). See Table 1 for basis of classification. (Based on data from OSPAR and EEA-member countries (Mediterranean), and data reported by Romania, see also Table 4). NB: larger symbols may obscure other symbols. 6

7 Map 2. DDT (expressed as DDE+DDD) in liver of cod (Gadus morhua) and muscle of herring (Clupea harrengus), median mg/kg wet weight for See Table 1 for basis of classification. NB: larger symbols may obscure other symbols. 7

8 Map 3. DDT (expressed as DDE+DDD) time trend in blue mussel (Mytilus edulis) in the north-east Atlantic (see Table 4). 8

9 Map 4. DDT (expressed as DDE+DDD) time trend in cod liver, herring liver/muscle, (see Table 5 6). 9

10 DATA Table 3. Relative changes (%) in DDT for concentrations in mussels (Mytilus edulis Me; M. galloprovincialis Mg), liver of Atlantic cod (Gadus morhua Gm), and muscle of herring (Clupea harengus Ch) (1990 = 100%). Sea Medium NE Atlantic Me NE Atlantic Gm Baltic Ch Table 4. DDT in mussels (Mytilus spp.), number of stations included in indicator calculations (median and/or trend). Mussels Atlantic Belgium France Germany Iceland Ireland Netherlands Norway Poland 2 Sweden Black Sea Romania 1 Table 5. DDT in cod (Gadus morhua), number of stations included in indicator calculations (median and/or trend) Cod Belgium Denmark 1 Finland 1 Iceland Norway Poland Sweden United Kingdom Table 6. DDT in herring (Clupea harengus), number of stations included in indicator calculations (median and/or trend) Herring Finland Iceland 1 1 Poland Sweden United Kingdom 3 More detailed information can be found in EEA file: HS-Fact-Sheet-DDT

11 Meta data Technical information 1. Data source No data on inputs of DDT is available. For assessment of concentrations in biota, OSPAR and Helcom data submitted via the ICES database to EEA provided data for the North-East Atlantic and Baltic. The assessment of concentrations in biota ( ) is based on unbalanced and inconsistent contributions from OSPAR, Helcom and EEA countries due to incomplete reporting to ICES (cf. Tables 4 6). For the supplementing data submitted by countries, more work is needed on ensuring reporting quality before they can be used in assessments. Supplementing data have also been received from other countries, but more work on quality assurance in the reporting of these data needs to be done before they can be used in the assessments. The issues relate in particular to incomplete meta-information about basis of measurements (wet/dry/lipid) and about observations below detection limit. There was insufficient data available to make a minimal assessment of levels and trends in sea water and sediment. DDTs are generally not monitored in seawater due to low water solubility. 2. Description of data Blue mussel data is from nine north-east Atlantic countries and one Black Sea country (Table 4 and Map 1). Atlantic cod data were available from eight north-east Atlantic countries (Table 5, Map 2). Herring data were available from five north-east Atlantic or Baltic countries (Table 6, Map 2). 3. Geographical coverage Data on concentrations in the blue mussel in the north-east Atlantic was available for every country bordering this sea area except Denmark and Portugal (Table 4). Data on concentrations in cod from the north-east Atlantic that were used are from Belgium, Denmark, Finland, Iceland, Norway, Poland, Sweden and the United Kingdom. (Table 5). Data for herring are from Finland, Iceland, Poland, Sweden and the United Kingdom. (Table 6). From newly independent States and eastern Europe in general, it has not been possible to include much data. Concentrations in mussels in 2001 have been reported from a single location on the Romanian coast, and these data have been included, although they do not necessarily give a representative picture of conditions in the Black Sea. 4. Temporal coverage Temporal trend data for biota generally covers the period , but most series are far from complete. Many series cover only a small part this period, for others, there are large gaps, up to nine years, of missing observations. In some areas, different stations were sampled in different years. The available data for each time series have been used for trend detection. The number of stations for each country and year are shown in Tables 4 6, and for a country varied from 0 to 25 stations monitored annually. For the Black Sea (Romania), the concentrations in mussel are only from Methodology and frequency of data collection For monitoring mussels and fish, member countries have applied different monitoring 11

12 guidelines. 6. Methodology of data manipulation Concentrations in mussels and fish Data for mussels and fish are reported on different bases (wet weight, dry weight, lipid weight), and with varying amount of information for converting between bases. The available data were inspected, and data were converted, using sample-specific information, to the basis that seemed to give the best coverage over stations and time, considered separately for species/tissue and parameter group. In cases where basis varies over time within station, data that could not be converted to the chosen basis from sample-specific information were discarded in the statistical analysis. Thus, it was chosen not to use data that had to be converted by general factors, as this might introduce artificial time trends in many cases. Some data sets were discarded because of insufficiently reported observations below detection limit that might bias the estimates. In such cases, whole data sets were discarded and not just below detection limit values. Time series were identified by station code where available, otherwise by location coordinates. Because of this, some data that constitute parts of the same time series may have been separated into different data series in the analysis because coordinates may vary for a station. Raw data were aggregated into yearly median values within each time series for the statistical analysis. The stations have varying amount of replicate samples per date, and varying number of dates per year. For each data series (species, tissue, location and contaminant), simple median values per year were calculated. For herring in the Baltic, log-transformed values have been used in the GLM analysis, and back-transformed in the chart. For all data, observations reported as below a defined detection limits were handled as ranges (0-detection limit), giving low-high ranges also for sums over parameters and for median values for several observations. Monotone time trends were tested using Mann-Kendall statistics (ICES, 1999), modified by taking low-high ranges explicitly into account in counting concordant, discordant and tied observation pairs. Time trends are based on all available data from 1985 to 1999 for each time series - the time coverage is very variable between series. However, some of the trends shown may be based on mainly older data. Significance of trends is based on two-sided test with a nominal 5% significance level, separately for each time series, without regard to serial correlation. Assessments of 'no trend' (i.e. no statistically significant trend) may be due both to actual lack of trend and to insufficient data. The median of the available data were calculated as measures of recent contaminant levels shown in the maps. Stations without any data after 1994 are not included in the maps. Medians were compared to Low and High concentrations (Table 1). The general changes in concentrations for each sea (Figure 1A and B) are based on the yearly aggregate values (averages) for each combination of location, year, species and tissue. Details of the calculation of these yearly averages are given in the text above. The overall time trends have been extracted from these aggregates for each species, tissue and region. Only series with data for at least three years, and with data at least up to 1994 have been used. To diminish the effect on apparent time trends of changing geographical coverage between years, overall yearly average values have been extracted by variance analysis (general linear model - GLM) with location and year as factors. The aim of this is to separate the variation due to stations from the change over time, and achieve yearly averages that are adjusted for 12

13 differences in geographical coverage between years. Some apparent changes in time may still be due to changes in geographical coverage between years. The analysis has been done using scaled values with the average value of each series, to get a representative average of relative variation over time, regardless of absolute levels at each station. Figures 1A and 1B show concentrations as a percentage of the average value for the year 1990, based on upper limits of estimated yearly aggregates. The regional trends only show the trends based on upper limits of yearly average ranges. The differences between lower limits and upper limits are large for all years before 1992 in Atlantic blue mussel. The lower limits still indicate higher levels before 1992 then afterwards, but the difference may be much smaller than shown in the figure. From 1992 to 1999 the lower-upper limits are quite close. For Baltic herring the differences between upper and lower limits are negligible. For Atlantic cod, the highest value in Figure 1B from 1997 may be somewhat inflated, but otherwise the differences are small between upper and lower limit estimates; the general picture, with fluctuations without any apparent long-term trend, will not be affected by this. Quality of information Lack of consistent or reliable data from the marine conventions or EEA member countries inhibit adequate assessment levels and trends of DDT in European marine waters. Aggregated data does not necessarily convey the uncertainty these problems cause, however, in general terms the more data that are used the more certain the results and the less sensitive the analysis is to missing data and unreliable data. 1. Strength and weakness (at data level): Some time series for biota are long with a fairly good QA, while others have incomplete temporal coverage, or lacks important information. Inadequate reporting of observation below detection limit is a problem in some data sets. This is particularly important for contaminants where delectability have improved over time due to analytical developments. Incomplete information about basis of measurement (wet weight, dry weight, lipid weight) is another important deficiency in some data sets. For some data sets, the basis has changed in time, without adequate information about conversion factors for each sample. This limits the use of the data for time trend analysis. 2. Reliability, accuracy, robustness, uncertainty (at data level): Concentrations in biota: For mussels and fish the quality of information depends on the choice of time series. The characteristics of this information depend on the choice of sampling strategy, completeness and duration of sampling, and the chemical analytical quality of data. For example, concentrations of hazardous substances in organisms may be sensitive to seasonal variations, age and sex of the organism (AMAP, 2000). Sampling strategies generally aim to control the influences of these variables and how well these strategies are followed are paramount to the quality of the results. Due to relatively long half-life times of some hazardous substances in organisms, the often slow change in expected exposure, and large natural variability, long time series may be needed to clearly determine trends. Temporal trend test: The Mann-Kendall test is a robust and widely accepted approach, but it has been applied here only on a series-by series basis. For independent observations within and between time series, about 5 % of trends will be deemed significant if in fact there is no trend. The 13

14 time series analysed here may have both serial correlation in time and spatial correlation between locations, so the overall significance of trends cannot without reservations be based on the frequency of significant individual trends. 3. Overall scoring (give 1 to 3 points: 1 = no major problems, 3 = major reservations): 3 Relevancy: poor (see below) Accuracy: poor no overall estimate of certainty of assessment insufficient investigation of serial of spatial correlation of data. No consideration of power or confidence limits in determining trends. Comparability over time: poor inconsistent sampling incomplete information Comparability over space: poor lacking knowledge of representativeness (hot spots versus locations representative of larger areas) Further work required (for data level and indicator level) Effort has to be made to provide input data at European level in general At the European level a monitoring strategy for load calculations should be developed leading to standardised and harmonised sampling and analysing procedures, and yearly reporting (on sampling conditions and on the representativeness of the samples). Apart from data completeness data quality should be given more attention. More effort has to be put into getting comparable and representative data on concentrations of hazardous substances in organisms from all European regional seas, with sufficient additional information about the data. The protocols and procedures for reporting and collecting data into Marinebase should be revised, based on the experience from the present work, and in consideration of the concerns discussed above, in collaboration with the various data sources. The procedures should ensure that sufficient information is given about reported data for a valid assessment of spatial and temporal development. More work should be done to identify data belonging to a single time series despite minor variations in coordinates. The methods for time trend assessments should be further developed, taking into account serial correlation in time, as well as spatial correlation between locations, to make regional assessments. A broader spectrum of methods should be considered (smoothers, parametric models). It should be investigated further how missing or incomplete data can be taken into account. Furthermore, there is a need at country level to report more intensive spatial and temporal trend monitoring of DDT. At the European level a monitoring strategy should be developed leading to standardised and harmonised sampling and chemical/statistical analytical procedures and yearly reporting (inter alia on sampling conditions and the representativeness of the samples) (ICES, 2000). Long-time series from fixed stations, representing the area or region in question, will be of great importance. Sampling the same stations each year will best ensure comparability among different years. Temporal trend assessment can be greatly improved by applying knowledge of local influences and consistent monitoring of representative stations where appropriate sampling 14

15 strategies have been applied. More information about the purpose and context of monitoring at the different locations would help in targeting the statistical analysis towards answering defined questions. For instance, locations strongly influenced by a point source, and with possible event-dominated variations in levels, should be considered separately from locations representing a general regional development. Development of management tools such as food limit values, background/reference concentrations (BCR) or ecotoxicological assessment criteria (EAC) would enhance and speed up the evaluation process of hazardous substances. References (more EU information can be found at AMAP (2000), Arctic Monitoring and Assessment Programme 2000, Arctic Pollution issues. Baan, P. J. A. and Groeneveld, G. J. J. (in press), Testing of indicators for the coastal and marine environment of Europe. Part 2: Hazardous substances, EEA Technical Report. Black Sea Commission (2002), State of the Environment of the Black Sea: Pressures and Trends , Ministerial meeting, Sofia, 14 June EU (2001), Decision No 2455/2001/EC of the European Parliament and the Council of 20 November 2001 establishing the list of priority substances in the field of water policy and amending Directive 2000/60/EC, 5 pp. Green, N. W., Knutzen, J (in press), Organohalogens and metals in marine fish and mussels and some relations to biological variables at reference localities in Norway, Marine Pollution Bulletin (submitted for review 30 April 2002). ICES (1999), ACME Report 1999, (see Chapter 6.3 and Annex 1). ICES (2000), Report of the Advisory Committee on the Marine Environment, ICES Cooperative Research Report No 239. Knutzen, J., 1987, Om Bakgrunnsnivåer av klorerte hydrokarboner og veslektede forbindelser i fisk (On background levels of organochlorines in fish), NIVA-report 2002, 173 pp. OSPAR (1998), OSPAR Strategy with regard to Hazardous Substances. OSPAR 98/14/1-E, Annex pp. OSPAR (1999), Report on assessment of trends in the concentrations of certain metals, PAHs and other organic compounds in the tissues of various fish species and blue mussels, OSPAR Ad Hoc Working Group on Monitoring pp plus appendices (see p. 10) 15